YouTube's crackdown on mass-produced, low-quality content generated by AI appears to be insufficient, with a new study finding that the platform's own recommendation algorithms continue to surface “AI slant” to new users.
According to a report published by video editing company Cupwing, more than 20% of videos suggested to new users by YouTube's algorithm are AI slops. The report analyzed 15,000 of the world's most popular YouTube channels to find out which channels are producing the most AI slop content, and how many views and revenue those videos generate.
Researchers found that 278 out of 15,000 YouTube channels uploaded only AI slop videos, garnering a total of 63 billion views and 221 million subscribers. According to Kapwing, the most viewed AI slop channel on YouTube (over 2.4 billion views) is called 'Bandar Apna Dost' and is based in India. The AI-generated video uploaded by this YouTube channel features an anthropomorphic rhesus macaque and another Hulk-like muscular character fighting demons.
Under YouTube's existing policies, AI slop videos are not eligible for monetization, but Kapwing estimates that such YouTube channels collectively generate about $117 million in annual revenue, with Bandar Apna Dost alone potentially generating about $4.25 million annually.
After creating a new account on YouTube, the researchers found that 104 of the first 500 videos recommended to their feed were AI slop, and another third could be classified as “brain rot.”
Kapwing’s findings provide a rare glimpse into a rapidly growing semi-structured industry that is using generative AI tools to increase engagement. It also includes another layer of people, mostly scammers, who offer paid tips and courses on how to create AI-generated viral content.
AI lag is particularly concerning because YouTube has a huge amount of authentic and inauthentic content, but can be mass-produced at scale using a suite of freely available AI tools.
Story continues below this ad
From Instagram to X to YouTube, social media users are increasingly complaining that their feeds are saturated with AI slop. In response, platforms are attempting to curb the spread of such low-quality content by tightening their policies and relying on removal mechanisms. Earlier this month, YouTube reportedly blocked two large channels selling fake AI-generated movie trailers.
But big tech companies also see AI-generated content as the future of social media. During an October earnings call, Meta CEO Mark Zuckerberg said Meta, the parent company of Facebook and Instagram, would be “adding an even larger corpus of content” to its recommendation system as AI “makes it easier to create and remix” the works shared online.
YouTube has also integrated Google's latest AI video generator, Veo 3, directly into Shorts, allowing users to create AI-generated vertical videos within the platform.
YouTube commented on Kapwing's report, saying, “Generative AI is a tool, and like any tool, it can be used to create both high-quality and low-quality content.” “We remain focused on connecting our users with high-quality content, regardless of how it was created. All content uploaded to YouTube must comply with our Community Guidelines, and we will remove content if it is found to violate our policies,” a company spokesperson was quoted as saying. Guardian.
Story continues below this ad
What is AI Slop?
Amidst the proliferation of AI-generated content on the internet over the past 12 months, American dictionary maker Merriam-Webster has announced that the word of the year for 2025 is slop. It defines the term as “low-quality digital content, typically produced in large quantities by artificial intelligence.”
“The flood of slops in 2025 included absurd videos, insane advertising images, cheap propaganda, fake news that looked pretty real, junk books written by AI, 'Workslop' reports that wasted co-workers' time…and lots of talking cats. People found it annoying, and people ate it up.”
“This is a very descriptive term. This is a part of AI, which is a transformative technology, that people have found fascinating, annoying, and a little bit ridiculous,” Greg Barlow, president of Merriam-Webster, was quoted as saying. Associated Press.
Other technology-related words have also been selected by dictionaries as Words of the Year. For example, Oxford Dictionaries selected “ragebait.” This refers to “online content that is intentionally designed to elicit anger or outrage by being irritating, provocative, or offensive, and is typically posted to increase traffic or engagement with a particular web page or social media content.”
Story continues below this ad
Collins Dictionary, on the other hand, has adopted “vibe coding”. This is a term coined by famous AI researcher Andrei Karpathy and refers to “the use of artificial intelligence prompted by natural language to write computer code.”
